Re: #645: Privacy impact of connection coalescing

> On 21 Nov 2014, at 4:49 pm, Greg Wilkins <gregw@intalio.com> wrote:
> 
> 
> On 21 November 2014 15:50, Mark Nottingham <mnot@mnot.net> wrote:
> The point is that these capabilities are new to this version of the protocol.
> 
> My point is that they are not really new.     HTTP/1.1 can be used to measure latency (302 will do it, or looking at TLS handshake round trips, connection:close also can give such info). HTTP/1.1 had persistent connections that can be used to associated between 12% to 100% of a clients requests depending on how many connections they use.
> 
> Furthermore, I think that such lists are only useful if they are complete enumerations.   The harm of calling them out explicitly is that it may give users/implementers the false impression that these are the only privacy concerns they need to worry about.

Noted.

--
Mark Nottingham   https://www.mnot.net/

Received on Friday, 21 November 2014 05:50:53 UTC